New AI video apps are doubling within weeks
The niche grows almost twice as fast as the market, and the apps moving fastest all shipped this year.
AI photo and video generators grew 7.5% at the median over ten weeks against 4% for the market. Of 1,302 apps we tracked in the niche, 197 at least doubled their installs, and 125 of those shipped in 2026. The entry cost is a model API and an app — no licence, no hardware, no partnerships.
Most fast-growing corners of the store are closed to a small team. Lending needs a licence, charging apps need physical stations, banking needs contracts. This one needs a developer.
What we see
The niche holds 2,414 apps that generate video or images: photo animators, AI music videos, avatar and face tools, clip makers. Together they account for about 5.8 billion installs.
Across the 1,302 of them we could measure at two points ten weeks apart, the median app grew 7.5%. The market median over the same window was 4%. That alone would be a mild signal. What makes it worth a second look is the distribution: 197 apps at least doubled their installs, and 125 of those were released in 2026.
The top of that list is steep. An app called Video Maker: AI Video Creator went from 11,000 installs to 540,000. AI MV Generator went from 14,000 to 500,000. Magic AI Photo Video Generator went from 169,000 to 2.9 million. All three shipped this year.
The median app in the niche sits at 35,000 installs. This isn't a market of giants — it's a market where a small app can still be visible.
What it might mean
The obvious reading is that generation models got cheap and good enough at roughly the same moment, and distribution on the store hasn't caught up with that yet. New entrants aren't out-engineering anyone; they're packaging a capability that was expensive eighteen months ago.
A less comfortable reading is that a lot of this is churn. Apps that grow tenfold in ten weeks on paid installs can shrink just as fast, and cumulative install counts never show the shrinking. We can see arrival, not retention.
A third possibility is that the category is being farmed: one team shipping several near-identical apps to see which one catches. The token profile of the niche — generator, maker, video, photo, face — is consistent with that.
What to check next
- Look at the release notes of the 2026 cohort, not their landing pages. Whoever is iterating weekly is testing something specific.
- Check review counts against installs. A tenfold install jump with flat reviews usually means paid traffic rather than pull.
- Find the underserved verb. "Generate a video" is crowded; the specific jobs around it — restoring old family photos, turning a product shot into a clip, making a lesson slide talk — are not equally crowded.
- Compare pricing pages. Where per-generation cost is passed straight to the user, margins are thin and the moat is nonexistent.
Limits
Ten weeks is two measurements, not a trend. We can say these apps had far more installs in August than in late May; we can't say the curve between those points was smooth, or that it continues.
Install counts are cumulative and never fall, so nothing here shows an app losing users. A niche can look healthy while every app inside it churns.
And the clustering is ours, not the store's: these 2,414 apps are grouped by what their listings describe, so the edges are debatable. A borderline photo editor with one generative feature may be in or out.
Figures come from our own dataset of the US Google Play catalogue, measured in late May and early August 2026.